On Labor Day, September 8, 2026, Motley Fool Hidden Gems Investing hosted a mailbag episode featuring guests Matt Frankel and Rachel Warren, who addressed three listener questions.
The first question, from Sabir in Austin, focused on investment metrics, specifically Return on Invested Capital (ROIC). Sabir noted that Coca-Cola had a higher ROIC than S&P Global (SPGI) and Waste Management (WM) but hadn't returned as much historically with dividends reinvested. He questioned if ROIC was the right metric and if comparing companies in different sectors was valid. Matt Frankel explained that ROIC is just one piece of the puzzle. Coca-Cola, being a massive and mature company, has fewer efficient places to reinvest its high returns into its business, leading it to distribute more as dividends, similar to Warren Buffett's See's Candy analogy. Rachel Warren agreed that comparing businesses across sectors can be valid for portfolio context, but ROIC is just one tool. She emphasized the importance of free cash flow (FCF) conversion, stating that high paper profits are meaningless if they don't convert to FCF for shareholders. Matt added that SPGI and WM had far more opportunities to reinvest—SPGI in data assets and new indices, WM in growing its footprint and recycling technologies—which compounded their intrinsic value and stock returns. Over 15 years, Coca-Cola saw a 327% total return (10.2% annualized), while WM returned 779% and SPGI over 1800%. Rachel also highlighted share count reduction through buybacks, funded by FCF, as a long-term compounding factor, but warned against debt-funded buybacks that swap dilution for financial risk. Qualitatively, Matt stressed pricing power (Coca-Cola, Waste Management) and essential services as competitive advantages. Rachel concluded that despite all metrics, valuation remains critical; even the best compounder can disappoint if an investor overpays.
The second question, from Isaiah, addressed the AI power bottleneck, comparing Enphase Energy's approach with Bloom Energy's. Isaiah noted AI data centers' growing electricity needs and asked how investors should compare Enphase's IQ solid-state transformer (targeting high-voltage DC architectures) with Bloom Energy's on-site power generation. Rachel framed the bottleneck as a two-part challenge: power generation and voltage conversion. Bloom Energy tackles generation by deploying on-site solid oxide fuel cells, allowing tech giants to bypass slow utility grid queues and get power quickly (90 days or less). Enphase, conversely, focuses on the conversion challenge, stepping voltage down efficiently at the chip level. Matt provided numbers for Enphase, estimating an 11 gigawatt U.S. market opportunity by 2031, with full system demos in late 2026, customer pilots in 2027, and volume shipments in 2028. He noted Enphase's core business was contracting (revenue down 20% year-over-year), making the data center opportunity a potential pivot. For Enphase's success, the 800-volt standard needs widespread adoption, and the company needs a hyperscaler win in 2027, alongside a stable residential business to fund these ambitions.
Regarding Bloom Energy, Rachel pointed to clear near-term monetization but also long-term risks, such as a potential slowdown in AI build-out, which could curb its significant growth. She also mentioned longer-term risks from shifting carbon mandates and fuel supply constraints as cleaner utility power scales. Matt affirmed that Bloom has a competitive moat, providing power much faster than alternatives, evidenced by a $20 billion backlog. He highlighted significant valuation risk, with Bloom trading at 80 times forward earnings and having surged 500% in one year. Matt summarized that Bloom is monetizing its opportunity *today*, while Enphase is a "2028 option on a shrinking company." For investors seeking exposure without betting on specific tech, Rachel suggested looking at physical grid infrastructure providers like Eaton and Schneider Electric, independent power producers, industrial storage providers, and companies supplying high-voltage cables and transformers, which benefit from multi-year backlogs regardless of specific chip architecture wins. The durable economics in the long term (5-10 years) are likely to be captured by physical infrastructure providers, energy asset owners, and land developers with regulated, contracted backlogs.
The final question, from Mike in Singapore, concerned the potential ripple effects of Anthropic and OpenAI IPOs, similar to what he observed with the SpaceX IPO impacting Alphabet and Rocket Lab. Matt clarified that the SpaceX IPO, the largest in history at the time, was extremely volatile in its initial weeks, causing sector rotation. Rocket Lab, for instance, saw significant drops around the IPO. This was largely "near-term noise" rather than a fundamental repricing. Rachel advised against attempting to time the market by selling holdings with the intention of buying back later, citing tax friction, execution costs, and the need to be "right two times." She noted that while SpaceX's IPO caused an industry-specific shakeup for Rocket Lab, the upcoming Anthropic and OpenAI IPOs might primarily affect big tech providers like Alphabet and Amazon, as institutional funds might trim legacy holdings to free up capital for pure-play AI. She suggested that any temporary pressure on Alphabet or Amazon could be a buying opportunity. Matt added that Anthropic is expected to go public first and potentially be much larger, targeting a $2 trillion-plus valuation. Alphabet and Amazon both hold stakes in Anthropic, which could lead to further sector rotation. Matt considers any weakness in Alphabet a buying opportunity. Both Matt and Rachel advised investors to focus on the financial health of the ecosystem, study public disclosures like the S1 filing, pay attention to lockup expirations, and remember any existing indirect exposure through holdings like Alphabet or Amazon.